HEURISTICS SUPPORTING USABLE AUTHORING TOOLS

Matching the right tool to the right user

Paula Kotzé, Elsabe Cloete

School of Computing, University of South Africa, Pretoria, South Africa

Keywords: E-learning, User needs, Functional and non-functional requirements, Interfaces and usability, Reusability,

Instructional authoring tools

Abstract: Over the past few years while e-learning has been gaining momentum, the user profile of instructional

authoring tools has also evolved. It seems that commercial authoring products have not yet been adapted to

address all user groups, which impedes lecturers in their working environment while preparing e-learning

materials, with the materials not achieving the required quality as a result. In this paper heuristics to design

an authoring tool aimed at a specific user group, namely the ordinary lecturer, are described to enable

subject-expert lecturers (not necessarily technically skilled) to create and reuse their own e-materials

without undergoing intensive technical training. The significance of these heuristics lies in the fact that they

provide a method to overcome many of the complexities associated with the design of instructional

authoring tools. Furthermore, tools developed according to these heuristics might enable institutions to cope

with the universal design demands associated with e-learning, without their e-learning programmes being

delayed by the scarcity of professional instructional designers and instructional programmers.

1 INTRODUCTION

Authoring support environments (ASEs) (also called

‘instructional authoring tools’) are mission-critical

applications in achieving usable e-learning

materials. Developing e-learning materials requires

competencies in several areas, including subject

matter, pedagogical foundations, instructional

systems development, and also ASE experience

(Tennyson, 2001). Cramer (2003) defines three

business application categories of ASEs that are

aimed at different user groups:

– ASEs complying with recognized instructional

design principles, requiring all the abovemen-

tioned competencies;

– ASEs that do not adhere to instructional design

practices, but create materials that may or may

not have a training focus; and

– entry-level ASEs offering restricted functionality

and limited deployment options without much

instructional design at the front end.

Ideally, educational institutions, and specifically

universities, should employ enough professional

instructional designers who can, together with

instructors, develop e-learning materials. In reality

the paucity of these professionals combined with the

potentially large number of courses (sometimes

thousands), often necessitate instructors having to

take responsibility for the development of part, or all

of their e-materials. Furthermore, some intermediate

communication/content/errata may be so diminutive

or urgent that it is not always possible to involve an

instructional design team in its composition, but the

lecturer is nevertheless obliged to handle its

introduction into the e-classroom. Since instructors

are often subject matter experts as well as being

trained in pedagogical approaches, we find that there

is a fourth ASE user group, namely those subject

experts who have educational foundations and

intuitive strategies to promote learning, but possibly

lack technical experience. This group requires an

efficient and usable ASE, simple in its approach but

without a corresponding loss of instructional quality

(Miller, 2003).

The question that we address in this paper is how

the usability of ASEs and the reusability of their

output could be improved in order to enable subject-

expert instructors, who are not necessarily

technically skilled, to create and reuse their own e-

materials without undergoing intensive technical

training.

Literature provides many guidelines on the

usability aspects of e-learning materials (see for

example IDE (2002), Macromedia (2002), Mehlen-

169

Kotzé P. and Cloete E. (2004).

HEURISTICS SUPPORTING USABLE AUTHORING TOOLS - Matching the right tool to the right user.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 169-178

DOI: 10.5220/0002612901690178

 SciTePress

bacher (2001), National Cancer Institute (n.d.), and

Step Two Designs (2002). However, design

guidelines to improve author usability in ASEs are

limited. The W3C (2000) recommends guidelines

for web authoring software developers to assist them

in designing ASEs that could produce accessible

web content, and also be accessible to their users.

These guidelines describe which usability issues are

to be addressed, while our discussion focuses on

how usability parameters can be introduced in ASEs.

Furthermore, the target user group of web authoring

tools assumes technical experience, which we cannot

assume for our target user group.

According to Duchastel (2001) information

design is an abstraction process determining what to

include and what to leave out; in effect determining

at what level and how to present specific content, in

this case, to the ASE author. The problem at hand is

complex due to:

– the number of ASE elements to consider;

– the fact that many of these elements are

amalgamated with the courseware elements,

which are aimed at a different target user group;

and

– the fact that many of these elements are

complex, being embedded in other elements.

What makes it complex to integrate usability

principles into ASEs, is the separation and isolation

of the amalgamated elements. One has only to

consider the various worldwide efforts to create

usable ASEs (based on open educational standards)

to become aware of the complexity of the situation

(De Vries, 2002). In this paper, we describe how to

disentangle the amalgamated elements that are

integral to both the authoring software and

courseware products. We believe by isolating the

elements it is possible to go beyond superficial

design issues and apply suitable usability parameters

to the isolated elements, before integrating them into

the authoring environment.

The significance of the heuristics that we have

established and report on in this paper, lies in their

provision of a method of overcoming many of the

complexities associated with the design of ASEs for

this specific target user group. Furthermore, a tool

that is developed according to these heuristics will

enable institutions to cope with the design demands

associated with e-learning, without their e-learning

programmes being stalled by the scarcity of

professional instructional designers. We do not

claim that the heuristics or the results are final or

complete. We have implemented them in a prototype

with limited functionality and share some of the

preliminary results. We recognize that further testing

needs to be done.

The entire e-learning domain forms an integrated

unit, and research focus on one level may easily fan

out to other levels, and direct or specify how these

levels should behave. Currently, there are several

standardisation efforts attempting to standardise

different aspects of the e-learning environment, as

this will improve coordination, reuse,

interoperability and vendor specifications

considerably. As the focus of our research is the

creation of courseware through a suitable tool, we

briefly refer to accomplishments that are focused on

the building blocks of courseware and their relation

to our work, before we commence the detailed

discussion of our work. The most prominent

international standardisation efforts for courseware

elements include IMS/EML (EML 2001, IMS 2002),

IEEE/LOM with Dublin Core (IEEE, n.d.), SCORM

(ADL, n.d.) and ARIADNE (n.d.). The purpose of

these standards is to describe information. For the

focus of our research, the purpose is to describe

learning objects so that they are independent,

reusable and easily retrievable units. The purpose of

our research is therefore focused on how to design a

lesson into which such units are integrated. There is

therefore a similarity in certain terminology, since

the basic building blocks are the same, yet there is

dissimilarity where our research focus swings away

from the building blocks to the method of creating

them.

Section 2 of this paper introduces a

characterisation of e-learning networks forming the

conceptual basis for the discussion that follows.

Section 3 is an exposition of the aforementioned

auxiliary

networks

auxiliary

networks

ASE glossary

network

IC glossary

network

ASE help

network

IC glossary

network

ASEN

ICN

ASE tutorial

network

ubnetworks

primary

networks

Figure 1: Internetwork of associated e-learning networks

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

170

design requirements. These requirements are used in

Section 4 to depict a design framework for ASEs

adhering to reusability and usability principles. In

the same section we discuss how we used this

framework to implement our ASE prototype.

Evaluation of the prototype is commented on in

Section 5, and conclusions are drawn in Section 6.

2 E-LEARNING NETWORKS

The e-learning environment can be considered as an

internetwork structure where different networks are

comprised of network segments, which in turn are

composed of associated nodes connected via links.

Nodes are functionally defined units representing

content. Categorising these nodes reduces the

complexity of the network. Typical category types

include glossary-type, help-type, question-type,

annotation-type, simulation-type, discussion-type,

browsing-type, and termination-type nodes (Kotzé,

1997). The names of these categories are self-

explanatory and as they suggest, these categories

provide a method of formalising content.

Links depict the potential routes between the

nodes within a network. As with nodes there are also

different link category types, including contextual,

referential, detour, annotational, return and terminal

links (Kotzé, 1997). Contextual links provide a

route-tracking mechanism recording the user’s path

through the network. Referential links are typically

the hyper-links that appear on an interface, allowing

the user to pursue different routes through the

network. Detour links connect the user to another

(usually auxiliary) network. Annotational links allow

the user to connect to an annotation node to record

personal notes. A return link returns from a

diversion initiated by a detour or annotational link,

and finally terminal links indicate that the end of the

current network has been reached.

Two primary networks are portrayed in Figure 1,

namely the Authoring Support Environment Network

(ASEN) and the Interactive Courseware Network

(ICN). The principal users of ASENs aim to create

courseware that produces an ICN for deployment on

CD, the Intranet or the Internet to be used by the

principle users of the ICN, namely the learners.

Figure 2 illustrates the input-output parameters of

the main networks, as well as their relation to one

another. The primary user interacts with the ASEN

by defining the different learning objects by using

plain text, graphics, defining links, et cetera. The

output of the ASEN then becomes the ICN, which is

used by both the primary ICN user (the learner) and

the secondary ICN user (the instructor).

The internetwork (illustrated in Figure 1)

consists of a set of states as well as relationships

between the states and allowable operations on these

states. We associate two states with each network;

namely an internal state and an external display

state. The internal state portrays the position or

condition of the nodes under development, as well as

the variables that are required to control interaction

with the author while creating and interacting with

these nodes. For example, while a course designer

creates a question-type node, values such as the type

of node being created, special attributes for the

particular node and conditions before the question or

parts thereof (like its answer) may be displayed, are

set. The external display state, on the other hand, is

the perceived network interface with which the actor

interacts. This state reflects parts of, but not

necessarily the entire internal state. For example,

the student who interacts with the question-type

event observes the content and certain conditions

that pertain to the question, but not the answer. The

completed ICN result state is exported for delivery

on a specific platform.

Figure 3, adapted from Kotzé (1997), illustrates

the mappings between the internal and external

display states within each network, as well as the

relationships among them. The interaction of the

actor with the ASEN nodes is reflected in mapping

(a) between the ASEN's internal and external states.

The internal state of the ASEN also maps to a

resulting ICN state in mapping (b) that shows how

the resulting ICN is reflected during the authoring

process. Furthermore, the resulting state of the ICN

ASE network

CREATE FORMAT

XML wrapped

objects

input

course

designer

IC network

output

student

interact

(learn)

interact

Figure 2: Input/output parameters of ASEN and ICN

HEURISTICS SUPPORTING USABLE AUTHORING TOOLS

171

has an externally perceivable rendering of the ICN

during the authoring process, as illustrated in (c).

Once the ICN has been exported to the required

delivery platform (during the learning process), the

internal state of the ICN maps to its external display

state illustrated in (d) (in the same way as for the

ASEN during the authoring process). The dotted

lines (i) and (ii) indicate that there is a relationship

between the different states in the ASEN and ICN.

As mentioned before, in this paper our focus and

interest lie in the ASE network. However, Figure 3

illustrates that it is impossible to consider the ASEN

in isolation since the states in this network have a

direct impact on the resulting ICN system.

There are also a number of auxiliary networks

attached to each of the primary networks to assist the

users in their interaction with the primary network.

For example, the ASEN in Figure 1 has three

auxiliary networks attached to it, namely the

glossary network, the help network and the tutorial

network. The user’s interaction with these networks

may enhance focus and efficiency, but is not an

essential element in achieving the user’s primary

goal. In fact, expert users seldom interact with these

types of networks. In many cases there are overlaps

between different auxiliary networks. These

overlaps are often non-deterministic, which means

that entrance from one network into another might

imply that a user could be trapped in a situation

where the network never terminates.

3 DESIGN REQUIREMENTS

Several sources were consulted and used to establish

tangible usability criteria for an ASEN interface

(ASTD, 2002; Badre, 2002; Cloete, 2003; Dix,

2004; Duchastel, 2001; Kotzé, 1997; Preece, 2002;

Proctor, 2002). These criteria are summarised below

and are helpful guiding principles to introduce

acceptable levels of usability in the ASEN interface

(ASENI). The criteria are organised according to the

basic principles of usability, namely learnability,

flexibility and robustness, and are then further

extended to the individual elements associated with

each.

Learnability is defined as the measure of how

easy it is to productively begin to use an interface

producing the desired results (Proctor, 2002).

Learnability elements of interest to the ASENI

include:

– Predictability: How consistent is the ASENI with

the author’s expectations?

– Familiarity: How well can the author relate the

ASENI interaction to his usual methods of

preparing learning materials?

– Generalisability: Is the degree of consistency

between the different ASENI elements high

enough to enhance the author’s predictability of

the interface?

– Internal consistency: Are there elements in the

ASENI that distract the author?

– Subjective satisfaction: Does the author like

using the interface?

Flexibility refers to the number of ways in which

the user can interact with the ASENI (Dix, 2004).

ASEN:

internal state

ASEN:

external display state

ICN:

delivered internal state

ICN:

external display state

(a)

(b)

(c)

(i)

(ii)

(relationship)

(mapping)

ICN:

resulting state

(d)

export

to delivery

platform

Figure 3: Mapping and relationships between ASEN and ICN

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

172

Usability parameters of interest that contribute to the

flexibility of the ASENI include:

– Dialogue initiative: Does pre-emptiveness lie

with the author or the ASEN?

– Multi-threading: Does the ASENI provide an

intuitive environment where the author can work

in multiple windows or on multiple tasks?

– Navigational functions: Can the author move

through the interface at own choosing?

– Task migratability: Can the author trust the ASE

to automate certain functions while taking

responsibility for others of his or her own

choosing?

– Customisability: Does the ASENI include

formatting, presentation, legibility options, as

well as WYSYWIG view capabilities?

– Orientation and tracking: Does the ASENI

include synthesisability features to track the

movements of the author, enabling him to

orientate himself with regard to the ASEN?

The robustness of a network refers to the

features supporting the successful achievement and

assessment of the goals (Dix, 2004). Usability

parameters affecting the robustness of an ASENI

include:

– Observability: To what degree can the author

readily determine the working of the ASE and

the interface?

– Error frequency, severity and recoverability:

What is the frequency with which the author

makes errors? How serious are these, and how

can they be recovered?

– Responsiveness: Does the ASENI give timely

feedback?

– Task conformance: To what degree does the

ASENI tasks comply with the intended actions

of the author?

4 DESIGN FRAMEWORK

The design decisions made during interaction with

the ASENI are pertinent to the ICN interface (ICNI).

Owing to this relationship between them, as

illustrated in Figures 1 to 3, it is not possible to

isolate the design criteria for the ASEN from the

design criteria of an ICN. Our proposed design

guidelines consist of four phases as set out in the

subsections to follow. However, before describing

each of these phases and how we implemented them,

we comment on the ASE prototype that we

developed according to these phases.

We developed an ASE prototype of limited

scope, meaning that the prototype covered only one

learning unit (see Section 4.1) as a subnetwork of

the primary ASEN. The objectives of this prototype

were twofold, namely to provide an authoring tool to

develop an ICN that is universally usable given the

target audience, as well as being able to create

learning objects that are based on open standards to

foster large-scale reuse. We have limited the scope

of our prototype to prove that the suggested

objectives are indeed achievable within such an

ASEN.

The prototype was developed so that on first-

time use, the author is prompted with the

opportunity to enter personal and general course

details such as course facilitator, course code and so

forth. Although this does not initially promote user

pre-emptiveness, it sets the scene for enhanced

orientation and tracking capabilities, adaptability,

and task migration.

4.1 Phase 1: Identify Learning Units

The initial focus of the design is on the content of

the learning situation, thus on the ICNI. The purpose

of Phase 1 is to provide the author with a tool to

identify manageable and sensible chunks of content,

which would typically be handled in one learning

session. These chunks are called learning units

(LUs). An example of LUs can be found when

considering a course, say Systems Analysis. Typical

LUs of this course might include basic concepts,

requirements gathering, requirements validation,

dynamic modelling, class modelling, and so on.

Defining LUs is therefore a fairly generic process,

not subject-specific, but merely a mechanism to

provide the author with the means to identify

different subject-specific chunks of content.

The term ‘learning unit’ correlates with the term

used in learning environments in general, which

increases familiarity when used in the ASENI. Once

the author has grasped the intended meaning of an

LU (it should actually be part of his/her pedagogical

foundation), constructing the courseware with LUs

bears a resemblance to designing a paper-based

lesson. Therefore, the use of LU elements in an

ASENI increases both the predictability and

generalisability of the interface because the author is

faced with a customary lesson design environment.

Implementation of LUs typically means the

inclusion of a menu option such as <Define new

LU>. This type of menu item (or interface button)

shifts the perceived pre-emptiveness from the

interface to the author. The basic interface layout of

our ASEN prototype, however, only contains

buttons to define, maintain or delete topics (see

Section 4.2) and events (see Section 4.3), since we

limited the scope of the prototype to include only

one LU, namely L

=COURSE INFORMATION.

HEURISTICS SUPPORTING USABLE AUTHORING TOOLS

173

Even though we created only one LU, we found

that users had an inherent understanding of the

concept of an LU. The specific articulation of

distinct LU elements (for example in a separate

window) and how they are linked together,

simplifies the inclusion of multi-threading and

navigational functions, whilst observability is

increased, with the author being able to perceive a

clear view of the course elements and hence the

course structure.

4.2 Phase 2: Topic Identification

Once again in this phase, we identify, isolate and

formalise ICN elements, but this time we focus on

the main topics within a specific LU. The focus in

this phase therefore remains on the content.

Returning to our previous example of the Systems

Analysis course and using the Basic Object-Oriented

Concepts LU, we can, for example, identify the main

topics in this phase as being objects and classes,

class attributes, object and class relationships,

methods, encapsulation, polymorphism, inheritance,

generalisation and specialisation, and so forth. Each

topic forms a network segment of the ICN.

Formally, each LU is comprised of a set of topics

that are epitomized by that particular LU:

As with the LUs, the articulation of topics in this

way increases predictability, familiarity,

generalisability, multi-threading, and the inclusion

of navigational functions. Furthermore, if the system

can be trusted to order the topics within the different

LUs, with the option that the user may change this

order by drag and drop activities, task migratability

can be promoted.

In our ASE prototype, we predefined a number

of specific topics. In general, topic selection should

not be content-specific, but since the topics of the

selected LU

(‘Course Information’) are universal

for all courses, we specified the following topics:

= OBJECTIVES T

=SYLLABUS

= PREREQUISITES T

=HOW_TO_STUDY

= MATERIAL T

=INTERNET_ACCESS

= COMMUNICATION T

=ASSIGNMENTS

= ASSESSMENT

4.3 Phase 3: Event Declaration

4.3.1 Main networks

During the third phase, the design focus of the ICN

merges with the design focus of the ASEN.

Teaching and learning occurs during the interaction

of various learning events such as questions, self-

tests, discussion sessions, and so forth. Our focus

therefore shifts to the methods conveying the

content. Our intention is to create a set of all

possible methods, and make them available to the

author, giving the author freedom to select one or

more events appropriate to render possible the

learning of a specific topic. This can be done by first

considering a set of events that are associated with

each topic, and then uniting all event sets so as to

capture the set that underpins the possible learning

methods. Each event becomes a node in the network

segment of the specific topic. We have defined such

a set of events as including:

E = {discussions, questions, examples, exercises, self-

tests, simulations, URIs} (URIs = Uniform Research

Identifiers, for example, a link to an external resource

such as a diagram, video clip, URL (Uniform

Research Locator – referring to a web address), et

cetera.)

ASENI designers can add to this set, but it is

advisable to guard against a too fine-grained

categorisation of events, as this might introduce

interaction complexities, especially where unfamiliar

terminology is used that focuses on computing

technologies and terminology rather than on learning

terminologies.

Each event has specific attributes making the

event flexible for use in different circumstances. In

the ICN, attributes play an important role in

constituting the learning environment, while in the

ASEN, the presentation of each attribute and how

the ASE user can interact with it, contributes to the

usability of the ASEN. To illustrate the point, we

briefly describe one type of event, the question-type

event, in more detail. For a detailed explanation of

all event descriptions, see Cloete and Kotzé (2003).

A question-type event is defined with

customisable attributes and can be constructed by

setting several attributes, of which only a few are

compulsory while the others default to NIL. We

define a question-type event as follows:

The first two attributes refer to the question

number and whether it is a main or sub-question –

allowing for task migratability, where the user can

either expect the system to handle the numbering or

use his/her own numbering scheme. The show

attribute is used to either display or withhold the

{

}

numbers natural are , k, where,

mjTLU

kjk

∃∀

{}

node.network

a as act can whichevent tindependen an form together that

events, other of composed event compound a E with

optValueoptQ and

QicrdhintansqstcmntshowsQuestionPonoQ

whereQEE

compound

jcompoundquest

,,,,,,,

∪=

∪

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

174

answer, depending on other criteria that the author

has. The author can add comments that are not

perceivable in the ICN by using the cmnt attribute.

The qst, ans, hint, and crd attributes refer to the

question itself, the answer, any hints that are to be

displayed in the ICNI, as well as the credits for the

question. For multiple-choice questions, the user can

activate the opt and optValue attributes. To enhance

customisability and learnability during

implementation, the network should initially respond

to the user’s request to create a question node by

producing a perceivable coherent interface window

where only compulsory attributes await input. A

<More Advanced..> button can then give the user

access to the other attributes. Keeping a counter to

determine how often the user interacts with the

<More Advanced..> button, the ASENI can adapt

the perceivable window to make the most frequently

accessed (or all) attributes available on the same

interface window as the compulsory attributes.

Owing to the limited scope of our ASE

prototype, we restricted the available events on the

interface to include discussion-type, exercise-type,

question-type, and link-type events. We briefly

explain how we associated these events with the

topics discussed in the previous phase. When the

author defines (selects) a new topic, one or more

events are either associated with it, or the author is

given the option to select specific events to associate

with the new topic. Whether this association is hard-

coded or created by the author, depends on the type

of topic. For example, the events associated with T

through to T

are hard coded as a combination of

discussion-type and link-type events since these

topics contain (flat) content that has to be presented

to the user, with little or no interaction expected

from the user. However, for T

(‘assignments’), the

author has the option of associating different, or a

combination of event types with each assignment

that is defined. As such, the author might start an

assignment (exercise-type event) by creating a

scenario (discussion-type event), followed by

references (link-type events), before stating the

problem (question-type event). Figure 4 represents a

screenshot from the ASEN prototype showing how

the exercise-type event is depicted.

4.3.2 Auxiliary networks

The structure of an auxiliary network is functionally

integrated with its primary network. Although it is

an autonomous network, it cannot be designed to be

entirely independent of, or in isolation from, its

primary network. We briefly mention three

archetypal nodes for the different auxiliary networks

and mention the related usability aspects. The

troubleshooting event assists one to find the reason

for inexplicable behaviour of the network and also to

find measures to improve it. A help event provides

an explanation of the purpose of another event or

terminology. A tutorial consists of several links

referring mainly to different discussion, simulation

and help events. The auxiliary networks thus add to

the robustness of the ASENI, specifically enhancing

observability and recoverability. Task conformance

is enhanced as help and simulation events can

explain the purpose of the task should the author

misunderstand it. We have not integrated any

auxiliary networks into our prototype at this stage.

4.4 Phase 4: Link plotting

After definition of the network nodes, network

routes are designed and specified in a navigational

table. In a simple situation, the navigational table

can be a simple indexed database. However, in a

fully developed network environment where the

primary networks are augmented with auxiliary

networks, is where a mesh of logical routes possibly

exists. In this case an improved store-and-retrieval

method is required, where entries in the navigational

table are stored in pairs of the format (s,n), where s

refers to the source node's ID and n refers to the next

node's ID. A next-hop routing algorithm uses a one-

step-along-the path approach to identify the next

node en route to the destination. As a first step in

determining a route from source to destination,

possible routes from the source are extracted. If none

of these provide a direct link to the destination,

entries that include the destination address are

extracted next, and their sources are followed

backwards until the shortest route from the source is

determined.

A design challenge is to provide a suitable tool

for route visualisation, route planning and route

creation. This can be achieved by creating an

environment where created nodes are displayed and

the author can visually connect nodes. Observability,

orientation and tracking are greatly enhanced by

making the completed mesh graph perceivable to the

user. However, for large routing tables, the complete

mesh may actually increase complexity of the

interface instead of simplifying it. In such a case,

flexibility parameters such as adaptability and

adaptivity should be given special attention during

the ASENI design, by making provision for the

interface to expose only a route cluster at a time,

instead of the complete mesh.

HEURISTICS SUPPORTING USABLE AUTHORING TOOLS

175

Navigation through the ASEN prototype is

straightforward, following the indexed database

approach.

4.5 Phase 5: Exporting and Delivery

The complicated part of developing the prototype

was to export the author’s content to an appropriate

standard output format after the authoring process.

The aim is also to export the learning objects to an

open standard that would enable reuse, without

burdening the user with the technical details of the

required standard.

We refrained from exporting the content during

the authoring process, as editing and multi-threading

capabilities would have complicated the

programming task tremendously.

The export transaction interacts with an XML

repository containing one-to-one mappings between

the components and their corresponding XML tags.

Each component has at least two tags associated

with it, namely a start-tag and a stop-tag. A first-in-

last-out method is used to pre-affix and append start-

and stop-tags to complex compositions of text and

meta-data. These components can be reused either

by including them in another LU, for example, or by

referencing them.

5 EVALUATION OF PROTOTYPE

The prototype was made available to a group of

seventy-five end-users belonging to the target user

group identified earlier. The majority of this group

were non-technologist teachers/instructors, and a

smaller number included instructional designers with

a technical background. Some members of this

group were previously exposed to an authoring tool

to code their learning units in EML (EML 2001,

IMS 2002). This required them to work directly with

the aforementioned XML editor. In general, this

group showed excitement and satisfaction at using

the prototype rather than working directly with

XML/EML tags.

Feedback and evaluation tests showed that the

users found our basic interface layout dealing with

topics and events intuitive, and strongly related this

layout to the paper-based preparation they were

Figure 4: ASENI prototype screenshot of Exercise environment

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

176

accustomed to. As anticipated, the users found the

layout to be predictable and familiar. They also

found it easy to generalise across different interface

functions. Notable usability features embedded in

the design of events include predictability and

familiarity, as the names of the event types are

exactly the terminology that instructors deal with

every day. We found that the way the design is

structured promoted author pre-emptiveness,

generalisability and task migratability. We were

largely satisfied that most of the learnability,

flexibility and robustness parameters were

addressed.

The following were identified as urgent

requirements for the next version of the prototype:

– The inclusion of an author analysis function

where the interface can sense the skills level of

the author and adjust the display and

functionality accordingly.

– The inclusion of a lesson preparation step

enabling the author to enter initial ideas when

planning a course, and also to be prompted with

suitable strategies and hints.

– Special attention will have to be paid to the

interface as the layout of this planning step can

easily compromise the usability of the entire

tool, should it require complex technical skills.

– Compliance with an open educational standard

such as DMCI / SCORM (IEEE, n.d.), or full

EML instead of a mere translation to XML tags.

6 CONCLUSION

Because an ASE network forms an intrinsic subset

of a set of e-learning networks, existing ASE

software requires expert users who are not

challenged by their computing design environments,

and are therefore able to focus their full attention on

the design of a learning environment. However,

lecturers are increasingly required to design e-

learning environments, and as a result are challenged

by the fact that their bags of professional skills do

not, by default, include computing skills and natural

software usage intuition. The implication of this

challenge is that lecturers struggle to focus their

attention on learning environment design, and are

impeded by the design environment.

In this paper we proposed an approach to

designing an ASE network that adheres to several of

the important usability parameters known in

software development, and at the same time

produces reusable, XML-wrapped output learning

objects. Our suggested approach of articulating

required ASE elements as network nodes enables the

designers to separate learning design issues from

interface usability issues. The design of the

prototype relied on the suggested methodology,

where different events to be included in the ASE

were designed through a set of formally defined

nodes.

Our prototype proved to overcome many of the

challenges that confront lecturers when they are

designing e-learning software. The prototype is

largely based on system pre-emptive dialogue

initiative, which impedes flexibility to a certain

extent. However, in the domain of authoring e-

learning software, the most important usability

design criteria are focussing on increasing usability

for novice or occasional users, rather than expert

users. A full implementation would obviously also

consider the expert user, and include a more user

pre-emptive approach. On the positive side, our

prototype greatly enhanced learnability, robustness

and most flexibility parameters.

The work reported upon in this article is partly

based on work sponsored by a grant from the

National Research Foundation of South Africa under

Grant Number GUN: 2050310.

REFERENCES

Advanced Distributed Learning (ADL), (n.d). Retrieved

April 2003 from http://www.adlnet.org.

ARIADNE, (n.d.). Retrieved April 2003 from

http://www.adlnet.org.

ASTD Certification Institute, 2002. ECC Standards.

Available from Retrieved October 2003 from

http://www.astd.org/ecertification/standards.htm.

Badre A.N., 2002. Shaping Web Usability. Addison

Wesley.

Cloete E. & Kotzé P., 2003. Interaction parameters in the

design of authoring support environments. Technical

Report: TR-UNISA-2003-01. Available from:

http://www.cs.unisa.ac.za/TechnicalReports/

National Cancer Institute, (n.d.). Usability Basics.

Retrieved October 2003 from

http://usability.gov/basics/index.html

Cramer G., (n.d.) Rapid Content Development Tools –

Selection Criteria. Retrieved April 2003 from

http://www.radauthor.com/pdf/Tools%20%Selection%

20Criteria.pdf.

De Vries, F., 2002. Valkenburg group moving forward.

Retrieved October 2003 from http://learningnetworks

.org/forums/

Dix A., Finley J., Abowd G.D., & Beale R., 2004. Human

Computer Interaction. Prentice Hall, Third Edition.

HEURISTICS SUPPORTING USABLE AUTHORING TOOLS

177

Duchastel P., 2001. Learnabililty. Retrieved October 2003

from http://home.earthlink.net/~castelnet/info/

learnability.htm.

EML, 2001. Open Universiteit Nederland - Learning

Networks. Retrieved November 2002 from

http://eml.ou.nl/eml-ou-nl.htm.

IDE, 2002. Innovations in Distance Education. Retrieved

October 2003 from http://www.outreach.psu.edu

/de/ide/.

IEEE P1484.12, (n.d). Learning object metadata working

group (IEEE’s LOM). Retrieved April 2003 from

http://ltsc.ieee.org/wg12/index.html.

IMS, (2002). IMS Global Learning Consortium. Retrieved

April 2003. Available at http://www.imsglobal.org.

Kotzé P., 1997. The Use of Formal Methods in the Design

of Interactive Authoring Support Environments. PhD

Thesis. Research Report YCST 97/09, Department of

Computer Science, University of York (UK).

Macromedia Inc., 2002. Macromedia Authorware Support

Center. Retrieved October 2003 from

http://www.macromedia.com/support/authorware.

Mehlenbacher B., 2001. Usable Web-based Instruction

Resources. Retrieved October 2003 from

http://www4.ncsu.edu:8030/~brad_m/publications.htm

l#Web-Based.

Miller M.G. & Cloete E., 2003. Authoring tool for

technical and technically-challenged e-learning

designers. In Proceedings of CATE 2003,

International Conference on Computers and Advanced

Technology in Education.

Preece J., Rogers Y., and Sharp H., 2002. Interaction

Design: Beyond Human Computer Interaction.

Addison Wesley.

Proctor R., 2002. Usability I: Principles and guidelines.

Retrieved October 2003 from http://www.informatics.

ed.ac.uk/teaching/modules/hci/slides6.pdf.

Step Two Designs Pty Ltd., 2002. How to evaluate a

content management system. KM Column. Retrieved

October 2003 from http://www.steptwo.com.au/

papers/kmc_evaluate/index.html.

Tennyson, R.D., 2001. Defining core competencies of an

instructional technologist. Computers in Human

Behavior, 17(4), 355-362.

W3C, 2000. Authoring Tool Accessibility Guidelines 1.0.

Retrieved October 2003 from http://www.w3.org/

TR/ATAG10/.

ICEIS 2004 - HUMAN-COMPUTER INTERACTION

178